AltInst: Re: Consensus Proportional Representation

David Chapman (<chapman@democdesignforum.demon.co.uk>) asked for
comments in June on

        http://www.democdesignforum.demon.co.uk/TwoESsWestm.html

which is a paper on two proposals for a new way of conducting an
election using proportional representation. I hope these comments are
still useful. My comments focus on the first proposal, called
Territorial Proportional Representation. The design is intended to
(1) strengthen each representative's ties to a particular geographic
constituency, and (2) avoid extremism (at least extremism that doesn't
reflect the electorate's views).

I found the proposal to be a very nice design. It allows each party
to divide the electorate into a number of regions and run a candidate
in each. It gives them an incentive to choose a number of regions
that reflects the number of seats they are likely to win. This gives
each candidate a territory that they will represent if they win, and
allows the party to lay out their strategy so they can choose a set of
territories they have a good chance of winning.

Some of the strengths of the proposal include:

* Candidates would be incented to pay less attention to opponents
individually, and more to their own and their opponents parties and
positions. This comes about because the different parties would
usually carve up their territories differently. This means that each
candidate, in each of the precints in which they run could be facing
different opponents from the same opposing party. If they attack a
single opponent on the basis of personality, only some of their
potential constituents would actually be voting for that opponent.
This leads to both less reliance on attacking the opposing
personality, and more emphasis on parties as consistent bundles of
policies. (the policy proposals would be more consistent from
candidate to candidate within the party, not that the policies
proposed by any candidate would be more consistent with one another.)

* Gerrymandering isn't a problem as far as I can tell. Each party
gets to draw their own district lines, and they can only move precints
among their own candidates, they can't trade one of their strong
precints for someone else's weak ones.

* The proposal allows more than one candidate from the same party to
run in the same territory. The votes are summed, and if there are
enough votes in that territory to send a representive, the one with
the most votes is elected. If there aren't enough votes, the party
can add the votes to those of other territories until there are enough
votes to earn a representive, and then send one from the entire
combined territory.

Some weaknesses:

* The combination rules seem to (intentionally) disenfranchise minor
parties. That is, smaller parties lose more to round-off errors that
larger parties. It was my impression that most proportional
representation systems were designed to enhance the strength of minor
parties above a certain threashold.

* Candidates know who they're running against, but face a different
variety of opponents in each constituency. If they are to respond to
the rhetoric of each opponent, the minor party candidates are at the
worst disadvantage. (They have few large territories, compared to the
major parties' many small territories.)

Chris

-- 
Chris Hibbert         It is easy to turn an aquarium into fish soup, but
hibbert@netcom.com    not so easy to turn fish soup back into an aquarium.
                        -- Lech Walesa on reverting to a market economy.
http://www.aimnet.com/~hibbert/home.html
PGP fingerprint = 1C 3B 30 DF 75 B2 18 F1 22 EF 11 18 0D 8A 3D A3
[To drop AltInst, tell:  majordomo@cco.caltech.edu  to: unsubscribe altinst]